each nearbAI core is an ultra-low power neural processing unit (NPU) and comes with an optimizer / neural network compiler. it provides immediate visual and spatial feedback based on sensory inputs, which is a necessity for live augmentation of the human senses.

optimized neural network inferencing for visual, spatial and other applications

unparalleled flexibility: customized & optimized for the customer’s use case

produces the most optimal NPU IP core for the customer’s use case: power, area, latency and memories trade-off

minimized development & integration time

ideal for battery-powered
mobile, XR and IoT devices

answers to your needs

ultra-low power

  • maximized battery performance
  • supports battery-powered always-on applications with wake-up
  • does not have to transport full video signals from the sensors as it supports local processing

zero-latency switching

  • continuous multiplexing between multiple neural networks, on the same nearbAI core
  • one neural network detects regions-of-interest and passes them to the next network, on the same nearbAI core, example: feed detected faces into face recognition network

local processing

  • each nearbAI operates standalone:
    • ultra-low latency
    • privacy conserving: local removal of sensitive data
    • high level of data security and reliability
  • nearbAI in datapath between sensors and XR processor
  • sensor fusion
  • supporting image sensors, ToF, LiDAR, microphones etc.
  • cloud interoperability (optional) with low system latency

scalable & flexible

  • optimized power-performance-area (PPA), based upon customer needs
  • silicon technology agnostic: 40nm down to 4nm and beyond
  • programmable using the optimizing compiler during the product lifecycle

optimizing nearbAI compiler

  • reads in your neural network in ONNX format using TensorFlow, Keras, PyTorch, mxnet
  • includes profiler with full layer-per-layer visibility
  • single-bit granularity data types

time-to-market acceleration

  • very fast FPGA prototyping in path towards ASIC
  • smooth integration of the easics IP core
  • efficient deployment and field upgrades via OTA

ultra-fast response time

  • live feedback within response time of human senses, from sensors to in-glass displays
  • real-time pattern recognition with fixed inference latency; face detection in 2 milliseconds

unique small footprint
/low cost

  • smallest silicon area
  • lowest ASIC cost and lowest PCB cost
  • no additional XR accelerator chips needed

easy SoC integration

  • application & integration templates available
  • supports Linux API & bare metal
  • peripheral and sensor interface blocks supported
  • support & integration team with +30 years experience / first-time right

secure OTA updates

  • field upgrades of neural network models and weights using secure SW over-the-air updates
  • lifecycle management
  • supports offload engine/processor outside IP core

function examples

3D / spatial
detection

face
detection

face
ID

face
blurring

iris
ID

gesture recognition
/hand tracking

object
detection

object
reconstruction

TXT

text
detection

IP cores & peripherals

⇧ click diagram for full-size view ⇧

why nearbAI?

highly computationally efficient and flexible NPUs

enable lightweight devices with long battery life ... with ultra-low power, run heavily optimized AI-based functions locally

enable truly immersive experiences ... achieve sensors-to-displays latency within the response time of the human senses

enable smart and flexible capabilities ... fill the gap between “swiss-army knife” XR / AI mobile processor chips and limited-capability edge IoT / AI chips

optimizer platform for ASIC

the L3 optimizer is a powerful ASIC optimizer platform that specializes in optimizing low power, low area, and low latency use cases for our customers. with the L3 optimizer, customers can expect customized solutions that are tailored to their specific needs, resulting in superior performance, reduced power consumption, and minimized area usage.

nearbAI is unique, delivering best-in-class solutions for battery-powered mobile, XR and IoT devices. It builds on more than three decades of DSP and image sensor chip design. Our mission is to provide customers with IPs that perfectly fit their specifications, offer outstanding performance, and are seamless to integrate right through to tape-out.

easics is committed to bringing the most advanced technology and solutions to market. With nearbAI, along with our partners and customers, we are pushing the latest AI innovation to the extreme edge
ramses valvekens
CSO & managing director
easics

delivery process

at easics, we believe in empowering designers by enabling them to realize their ideas, test their practicality, and gather user feedback.

delivery options incl. customization

we are very flexible, so you can choose the solutions that best fit your needs. 

let’s do a custom benchmark together

provide us with your use case:

quantized or unquantized NN model(s):
ONNX, TensorFlow (Lite), PyTorch, or Keras

constraints:
average power & energy per inference, silicon area, latency, memories, frame rate, image resolution, foundry + technology node

first-time right

we are also offering:

Subscribe to our newsletter

Don't miss new updates in your email